Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
@@ -1,39 +1,29 @@
|
|
1 |
import os
|
2 |
import gradio as gr
|
3 |
import requests
|
4 |
-
import inspect
|
5 |
import pandas as pd
|
6 |
-
from huggingface_hub import InferenceClient
|
7 |
-
from transformers import AutoTokenizer # Import AutoTokenizer from transformers
|
8 |
-
import chess
|
9 |
-
import chess.engine
|
10 |
-
from PIL import Image
|
11 |
-
from io import BytesIO
|
12 |
-
import base64
|
13 |
-
import re
|
14 |
from huggingface_hub import login
|
15 |
-
|
16 |
|
17 |
# --- Constants ---
|
18 |
DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
|
19 |
|
20 |
# --- Basic Agent Definition ---
|
21 |
-
|
22 |
class BasicAgent:
|
23 |
def __init__(self):
|
24 |
print("BasicAgent initialized.")
|
25 |
-
|
26 |
-
|
27 |
-
|
28 |
-
|
29 |
-
|
30 |
-
|
31 |
-
|
32 |
-
|
33 |
-
|
34 |
-
|
35 |
-
|
36 |
-
|
37 |
|
38 |
def maybe_reversed(self, text: str) -> bool:
|
39 |
words = text.split()
|
@@ -46,9 +36,9 @@ class BasicAgent:
|
|
46 |
return reversed_ratio > 0.3
|
47 |
|
48 |
def solve_riddle(self, question: str) -> str:
|
49 |
-
|
50 |
-
if "opposite of the word" in
|
51 |
-
match = re.search(r"opposite of the word ['\"](\w+)['\"]",
|
52 |
if match:
|
53 |
word = match.group(1).lower()
|
54 |
opposites = {
|
@@ -65,17 +55,14 @@ class BasicAgent:
|
|
65 |
print("Detected likely reversed riddle.")
|
66 |
return self.solve_riddle(question)
|
67 |
return "FINAL ANSWER: NOT_A_RIDDLE"
|
68 |
-
|
69 |
-
|
70 |
-
|
71 |
-
|
72 |
-
"""
|
73 |
-
# --- Determine HF Space Runtime URL and Repo URL ---
|
74 |
-
space_id = os.getenv("SPACE_ID") # Get the SPACE_ID for sending link to the code
|
75 |
|
76 |
if profile:
|
77 |
-
username= f"{profile.username}"
|
78 |
-
print(
|
79 |
else:
|
80 |
print("User not logged in.")
|
81 |
return "Please Login to Hugging Face with the button.", None
|
@@ -84,66 +71,47 @@ class BasicAgent:
|
|
84 |
questions_url = f"{api_url}/questions"
|
85 |
submit_url = f"{api_url}/submit"
|
86 |
|
87 |
-
# 1. Instantiate Agent ( modify this part to create your agent)
|
88 |
try:
|
89 |
agent = BasicAgent()
|
90 |
except Exception as e:
|
91 |
print(f"Error instantiating agent: {e}")
|
92 |
return f"Error initializing agent: {e}", None
|
93 |
-
|
94 |
agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"
|
95 |
print(agent_code)
|
96 |
|
97 |
-
# 2. Fetch Questions
|
98 |
-
print(f"Fetching questions from: {questions_url}")
|
99 |
try:
|
100 |
response = requests.get(questions_url, timeout=15)
|
101 |
response.raise_for_status()
|
102 |
questions_data = response.json()
|
103 |
if not questions_data:
|
104 |
-
|
105 |
-
return "Fetched questions list is empty or invalid format.", None
|
106 |
-
print(f"Fetched {len(questions_data)} questions.")
|
107 |
except requests.exceptions.RequestException as e:
|
108 |
-
print(f"Error fetching questions: {e}")
|
109 |
return f"Error fetching questions: {e}", None
|
110 |
-
except requests.exceptions.JSONDecodeError as e:
|
111 |
-
print(f"Error decoding JSON response from questions endpoint: {e}")
|
112 |
-
print(f"Response text: {response.text[:500]}")
|
113 |
-
return f"Error decoding server response for questions: {e}", None
|
114 |
-
except Exception as e:
|
115 |
-
print(f"An unexpected error occurred fetching questions: {e}")
|
116 |
-
return f"An unexpected error occurred fetching questions: {e}", None
|
117 |
|
118 |
-
# 3. Run your Agent
|
119 |
results_log = []
|
120 |
answers_payload = []
|
121 |
-
print(f"Running agent on {len(questions_data)} questions...")
|
122 |
for item in questions_data:
|
123 |
task_id = item.get("task_id")
|
124 |
question_text = item.get("question")
|
125 |
if not task_id or question_text is None:
|
126 |
-
print(f"Skipping item with missing task_id or question: {item}")
|
127 |
continue
|
128 |
try:
|
129 |
submitted_answer = agent(question_text)
|
130 |
answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer})
|
131 |
results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
|
132 |
except Exception as e:
|
133 |
-
|
134 |
-
results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": f"AGENT ERROR: {e}"})
|
135 |
|
136 |
if not answers_payload:
|
137 |
-
print("Agent did not produce any answers to submit.")
|
138 |
return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)
|
139 |
|
140 |
-
|
141 |
-
|
142 |
-
|
143 |
-
|
|
|
144 |
|
145 |
-
# 5. Submit
|
146 |
-
print(f"Submitting {len(answers_payload)} answers to: {submit_url}")
|
147 |
try:
|
148 |
response = requests.post(submit_url, json=submission_data, timeout=60)
|
149 |
response.raise_for_status()
|
@@ -155,59 +123,17 @@ class BasicAgent:
|
|
155 |
f"({result_data.get('correct_count', '?')}/{result_data.get('total_attempted', '?')} correct)\n"
|
156 |
f"Message: {result_data.get('message', 'No message received.')}"
|
157 |
)
|
158 |
-
|
159 |
-
results_df = pd.DataFrame(results_log)
|
160 |
-
return final_status, results_df
|
161 |
-
except requests.exceptions.HTTPError as e:
|
162 |
-
error_detail = f"Server responded with status {e.response.status_code}."
|
163 |
-
try:
|
164 |
-
error_json = e.response.json()
|
165 |
-
error_detail += f" Detail: {error_json.get('detail', e.response.text)}"
|
166 |
-
except requests.exceptions.JSONDecodeError:
|
167 |
-
error_detail += f" Response: {e.response.text[:500]}"
|
168 |
-
status_message = f"Submission Failed: {error_detail}"
|
169 |
-
print(status_message)
|
170 |
-
results_df = pd.DataFrame(results_log)
|
171 |
-
return status_message, results_df
|
172 |
-
except requests.exceptions.Timeout:
|
173 |
-
status_message = "Submission Failed: The request timed out."
|
174 |
-
print(status_message)
|
175 |
-
results_df = pd.DataFrame(results_log)
|
176 |
-
return status_message, results_df
|
177 |
-
except requests.exceptions.RequestException as e:
|
178 |
-
status_message = f"Submission Failed: Network error - {e}"
|
179 |
-
print(status_message)
|
180 |
-
results_df = pd.DataFrame(results_log)
|
181 |
-
return status_message, results_df
|
182 |
except Exception as e:
|
183 |
-
|
184 |
-
print(status_message)
|
185 |
-
results_df = pd.DataFrame(results_log)
|
186 |
-
return status_message, results_df
|
187 |
|
188 |
|
189 |
-
# --- Build Gradio Interface
|
190 |
with gr.Blocks() as demo:
|
191 |
gr.Markdown("# Basic Agent Evaluation Runner")
|
192 |
-
gr.Markdown(
|
193 |
-
"""
|
194 |
-
**Instructions:**
|
195 |
-
1. Please clone this space, then modify the code to define your agent's logic, the tools, the necessary packages, etc ...
|
196 |
-
2. Log in to your Hugging Face account using the button below. This uses your HF username for submission.
|
197 |
-
3. Click 'Run Evaluation & Submit All Answers' to fetch questions, run your agent, submit answers, and see the score.
|
198 |
-
---
|
199 |
-
**Disclaimers:**
|
200 |
-
Once clicking on the "submit button, it can take quite some time ( this is the time for the agent to go through all the questions).
|
201 |
-
This space provides a basic setup and is intentionally sub-optimal to encourage you to develop your own, more robust solution. For instance for the delay process of the submit button, a solution could be to cache the answers and submit in a seperate action or even to answer the questions in async.
|
202 |
-
"""
|
203 |
-
)
|
204 |
-
|
205 |
gr.LoginButton()
|
206 |
-
|
207 |
run_button = gr.Button("Run Evaluation & Submit All Answers")
|
208 |
-
|
209 |
status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False)
|
210 |
-
# Removed max_rows=10 from DataFrame constructor
|
211 |
results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True)
|
212 |
|
213 |
run_button.click(
|
@@ -216,25 +142,5 @@ with gr.Blocks() as demo:
|
|
216 |
)
|
217 |
|
218 |
if __name__ == "__main__":
|
219 |
-
print("\n" + "-"*30 + " App Starting " + "-"*30)
|
220 |
-
# Check for SPACE_HOST and SPACE_ID at startup for information
|
221 |
-
space_host_startup = os.getenv("SPACE_HOST")
|
222 |
-
space_id_startup = os.getenv("SPACE_ID") # Get SPACE_ID at startup
|
223 |
-
|
224 |
-
if space_host_startup:
|
225 |
-
print(f"✅ SPACE_HOST found: {space_host_startup}")
|
226 |
-
print(f" Runtime URL should be: https://{space_host_startup}.hf.space")
|
227 |
-
else:
|
228 |
-
print("ℹ️ SPACE_HOST environment variable not found (running locally?).")
|
229 |
-
|
230 |
-
if space_id_startup: # Print repo URLs if SPACE_ID is found
|
231 |
-
print(f"✅ SPACE_ID found: {space_id_startup}")
|
232 |
-
print(f" Repo URL: https://huggingface.co/spaces/{space_id_startup}")
|
233 |
-
print(f" Repo Tree URL: https://huggingface.co/spaces/{space_id_startup}/tree/main")
|
234 |
-
else:
|
235 |
-
print("ℹ️ SPACE_ID environment variable not found (running locally?). Repo URL cannot be determined.")
|
236 |
-
|
237 |
-
print("-"*(60 + len(" App Starting ")) + "\n")
|
238 |
-
|
239 |
print("Launching Gradio Interface for Basic Agent Evaluation...")
|
240 |
demo.launch(debug=True, share=False)
|
|
|
1 |
import os
|
2 |
import gradio as gr
|
3 |
import requests
|
|
|
4 |
import pandas as pd
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
from huggingface_hub import login
|
6 |
+
import re
|
7 |
|
8 |
# --- Constants ---
|
9 |
DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
|
10 |
|
11 |
# --- Basic Agent Definition ---
|
|
|
12 |
class BasicAgent:
|
13 |
def __init__(self):
|
14 |
print("BasicAgent initialized.")
|
15 |
+
self.agent_prompt = (
|
16 |
+
"""You are a general AI assistant. I will ask you a question. Report your thoughts, and
|
17 |
+
finish your answer with the following template: FINAL ANSWER: [YOUR FINAL ANSWER].
|
18 |
+
YOUR FINAL ANSWER should be a number OR as few words as possible OR a comma separated
|
19 |
+
list of numbers and/or strings.
|
20 |
+
If you are asked for a number, don't use comma to write your number neither use units such as $
|
21 |
+
or percent sign unless specified otherwise.
|
22 |
+
If you are asked for a string, don't use articles, neither abbreviations (e.g. for cities), and write the
|
23 |
+
digits in plain text unless specified otherwise.
|
24 |
+
If you are asked for a comma separated list, apply the above rules depending of whether the element
|
25 |
+
to be put in the list is a number or a string."""
|
26 |
+
)
|
27 |
|
28 |
def maybe_reversed(self, text: str) -> bool:
|
29 |
words = text.split()
|
|
|
36 |
return reversed_ratio > 0.3
|
37 |
|
38 |
def solve_riddle(self, question: str) -> str:
|
39 |
+
question = question[::-1] # properly reverse before parsing
|
40 |
+
if "opposite of the word" in question:
|
41 |
+
match = re.search(r"opposite of the word ['\"](\w+)['\"]", question)
|
42 |
if match:
|
43 |
word = match.group(1).lower()
|
44 |
opposites = {
|
|
|
55 |
print("Detected likely reversed riddle.")
|
56 |
return self.solve_riddle(question)
|
57 |
return "FINAL ANSWER: NOT_A_RIDDLE"
|
58 |
+
|
59 |
+
|
60 |
+
def run_and_submit_all(profile: gr.OAuthProfile | None):
|
61 |
+
space_id = os.getenv("SPACE_ID")
|
|
|
|
|
|
|
62 |
|
63 |
if profile:
|
64 |
+
username = f"{profile.username}"
|
65 |
+
print("User logged in.")
|
66 |
else:
|
67 |
print("User not logged in.")
|
68 |
return "Please Login to Hugging Face with the button.", None
|
|
|
71 |
questions_url = f"{api_url}/questions"
|
72 |
submit_url = f"{api_url}/submit"
|
73 |
|
|
|
74 |
try:
|
75 |
agent = BasicAgent()
|
76 |
except Exception as e:
|
77 |
print(f"Error instantiating agent: {e}")
|
78 |
return f"Error initializing agent: {e}", None
|
79 |
+
|
80 |
agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"
|
81 |
print(agent_code)
|
82 |
|
|
|
|
|
83 |
try:
|
84 |
response = requests.get(questions_url, timeout=15)
|
85 |
response.raise_for_status()
|
86 |
questions_data = response.json()
|
87 |
if not questions_data:
|
88 |
+
return "Fetched questions list is empty or invalid format.", None
|
|
|
|
|
89 |
except requests.exceptions.RequestException as e:
|
|
|
90 |
return f"Error fetching questions: {e}", None
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
91 |
|
|
|
92 |
results_log = []
|
93 |
answers_payload = []
|
|
|
94 |
for item in questions_data:
|
95 |
task_id = item.get("task_id")
|
96 |
question_text = item.get("question")
|
97 |
if not task_id or question_text is None:
|
|
|
98 |
continue
|
99 |
try:
|
100 |
submitted_answer = agent(question_text)
|
101 |
answers_payload.append({"task_id": task_id, "submitted_answer": submitted_answer})
|
102 |
results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": submitted_answer})
|
103 |
except Exception as e:
|
104 |
+
results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": f"AGENT ERROR: {e}"})
|
|
|
105 |
|
106 |
if not answers_payload:
|
|
|
107 |
return "Agent did not produce any answers to submit.", pd.DataFrame(results_log)
|
108 |
|
109 |
+
submission_data = {
|
110 |
+
"username": username.strip(),
|
111 |
+
"agent_code": agent_code,
|
112 |
+
"answers": answers_payload
|
113 |
+
}
|
114 |
|
|
|
|
|
115 |
try:
|
116 |
response = requests.post(submit_url, json=submission_data, timeout=60)
|
117 |
response.raise_for_status()
|
|
|
123 |
f"({result_data.get('correct_count', '?')}/{result_data.get('total_attempted', '?')} correct)\n"
|
124 |
f"Message: {result_data.get('message', 'No message received.')}"
|
125 |
)
|
126 |
+
return final_status, pd.DataFrame(results_log)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
127 |
except Exception as e:
|
128 |
+
return f"Submission Failed: {e}", pd.DataFrame(results_log)
|
|
|
|
|
|
|
129 |
|
130 |
|
131 |
+
# --- Build Gradio Interface ---
|
132 |
with gr.Blocks() as demo:
|
133 |
gr.Markdown("# Basic Agent Evaluation Runner")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
134 |
gr.LoginButton()
|
|
|
135 |
run_button = gr.Button("Run Evaluation & Submit All Answers")
|
|
|
136 |
status_output = gr.Textbox(label="Run Status / Submission Result", lines=5, interactive=False)
|
|
|
137 |
results_table = gr.DataFrame(label="Questions and Agent Answers", wrap=True)
|
138 |
|
139 |
run_button.click(
|
|
|
142 |
)
|
143 |
|
144 |
if __name__ == "__main__":
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
145 |
print("Launching Gradio Interface for Basic Agent Evaluation...")
|
146 |
demo.launch(debug=True, share=False)
|